We are studying the control mechanisms of the expression of the gal operon of E. coli. We have demonstrated that the operon is controlled by two promoters, which are midulated by cyclic AMP in opposite ways. Both the promoters are negatively regulated by a gal repressor protein. We have also previously shown that each of the two gal promoters is negatively regulated by two operator elements; one of which (0E) is located upstream to the promoters and the other (0T) inside the galE structural gene. The existence of the two operators, which do not overlap with the promoters, have now been shown by three ways: (i) Genetic and DNA sequence studies of operator mutations; (ii) Repressor-operator interactions with purified repressor protein and operator DNA fragments; and (iii) Chemical protection of the operator DNA segments by repressor from DNase digestion and phosphate ethylation. These results also show that repressor binds to one and the same side of the two operators. The repressor does not compete with cAMP-CRP or RNA polymerase to bind to gal DNA. 0E and 0I are separated by 96 bp. Insertion of a 15 bp fragment, which introduces an estra half of a DNA double helical turn between them causes derepression of the operon. Taken together, our results suggest that gal repressor does not sterically hinder the binding of the other proteins to ensue repression. We have proposed that repressor molecules bound to 0E and 0I interact with each other to generate a DNA loop, which includes the promoters. This changes the conformation of the DNA structure, thus making it inadequate for promoter activity.